Understanding trends and variation in paediatric fracture management in England

In this report we have taken paediatric Emergency department (ED) attendances (0-16yrs old) across England (2019-23) and looked at a select number of paediatric fractures where either national guidance or published/emerging evidence suggests limited intervention, imaging or follow up is required. These include clavicle (collarbone), forearm and toe fractures. Given our results we believe future UK wide analysis of other fracture types and/or a wider age range (Adults) may prove beneficial.

For paediatric fractures regarded as minor or ‘low need’, our analysis demonstrates:

• If the number of follow up appointments was reduced to the lowest quartile of English Trusts circa. 30,000 appointments could be saved annually. (Slide 40 & 49)

• Many clavicle fractures, commonly accepted to heal without intervention, have two or more follow up appointments when evidence suggests none are required. (Slide 23 & 41)

• Marked variance exists across Trusts in the proportion of fractures with follow up and how that occurs (virtual vs face to face). (Slide 39-40)

Our analysis demonstrates between 2019-23 the proportion of forearm fractures manipulated:

• Has reduced in theatre by 50% (Slide 27) and increased in ED (Slide 26).

• Total manipulations (theatre or ED) have reduced by 22% (Slide 28).

Our data shows that forearm fractures are the most common fracture in children and in 2022-23 alone over 25,000 had no follow up (Slide 23). Along with the forearm manipulation data this collectively suggests a fundamental change in forearm fracture management has occurred across England, likely driven by Get It Right First Time (GIRFT) targets and recent/ongoing research (FORCE & CRAFFT).

Lastly, our analysis suggests marked regional variation in fracture incidence exists (Slide 13) and is unaccounted in the current GIRFT/Model hospital ‘Wrist manipulations in theatre’ dataset. We have formulated an alternative and potentially more nuanced version accounting for regional variation. Interestingly, Trusts currently considered either in the top or bottom quartiles markedly change position within our proposed model (Slide 43-44).

In conclusion we believe our analysis can help set future GIRFT targets driving change and savings across the NHS.

Dr Patrick Aldridge, Consultant Paediatrician, Paediatric Emergency Department Clinical Lead at NHS Frimley Health Foundation Trust

One hospital in Ireland calculated a paediatric fracture incidence rate of approximately 29 fractures/1,000/year¹. The most common fracture was distal radial/buckle fractures (27.2%), followed by distal humerus /supracondylar fracture (13.9%), ankle fractures (9.2%), phalanx fractures (8.3%), and radial/ulnar metaphysis fractures (5.7%). It is suggested the incidence rate will depend on the social emphasis on encouraging physical activity¹.

Between 2012–2019, 368,120 children under 18 were admitted to English NHS hospitals with a fracture; 256,008 (69.5%) were upper limb and 85,737 (23.3%) were lower limb fractures². The annual incidence of upper limb fractures was highest in children aged 5–9 (348.3 per 100, 000 children).

The total fracture incidence rate in the UK was found by one study to be 1,370 fractures per 100,000 children³, while another found the incidence of distal radius fractures to be 337 fractures per 100,000 children at one UK trauma centre⁴.

British Society for Children’s Orthopaedic Surgery (BSCOS) guidance suggests that no referral/follow up is required for many fractures of the clavicle, elbow, wrist and toes where there is no or minimal displacement⁵.

The FORCE study found in children with a torus fracture of the distal radius there was no difference in outcomes between those who were offered of a bandage and immediate discharge (as per UK National Institute for Health and Clinical Excellence recommendations) and those receiving current treatment of rigid immobilisation and follow-up⁶.

One study in Scotland found that uncomplicated paediatric clavicle fractures can be managed without x-rays in the ED as the use of x-rays did not influence fracture management or add valuable information to clinicians’ assessment⁷.

A GIRFT report on Paediatric Trauma and Orthopaedic Surgery found that over 250 weeks of theatre time a year had been used for manipulation of forearm fractures between 2016 and 2019¹. A significant proportion of these displaced or angulated wrist fractures could have been manipulated and cast in the emergency department rather than being admitted and treated in the operating theatre. The GIRFT report found significant variation between trusts in the number of manipulations being performed in theatre. If the number of fractures manipulated in theatre at all trusts was reduced to the level seen in trusts with well-developed emergency department manipulation protocols there would be an 80% reduction in manipulations in theatre, reducing theatre time for forearm/wrist fractures to under 57 weeks.

Due to pressures on hospitals from the COVID-19 pandemic the British Orthopaedic Association developed guidelines for the early management of distal forearm fractures in children. A study conducted at one trust found that the implementation of these guidelines resulted in 86% of distal forearm fractures in children being manipulated in the emergency department, an increase from 32% prior to the COVID pandemic². This saved approximately 63 hours of theatre time in the six-month study period.

The GIRFT report did highlight reluctance to perform procedures in the emergency department because of worries about breaching the 4 hour emergency department treatment target, a lack of space/facilities to perform sedation and a lack of familiarity with techniques¹.

The study population included all those aged 16 and under who attended an emergency care centre in England between April 2019 and March 2024 and had a SNOMED code for a closed fracture of toe, clavicle, elbow, forearm or tibia/fibula recorded (see Appendices A-C for full code lists).

Records at a patient level were taken from the Emergency Care dataset (ECDS) and linked with any subsequent fracture related activity recorded in the Outpatient (OPA) and Admitted Patient Care Episode (APCE) datasets in the 3 months post-emergency care attendance. All data were accessed through the National Commissioning Data Repository (NCDR).

This study focuses on closed fractures as this is where there is scope for more conservative management; open, pathological, osteoporotic and birth trauma fractures have been excluded, alongside fractures of the great toe, which should all be followed up.

Incidence rates were calculated using the Office of National Statistics (ONS) mid-year population estimates by age and sex¹.

The coding of fractures is not sufficiently detailed and reliable to determine specific fracture types, and thus what would be the appropriate treatment at an individual patient level. For example, we cannot identify buckle fractures, which we know could be treated more conservatively. However, we could calculate the proportion of children with each fracture type that:

• had a fracture manipulated in theatre (only includes closed manipulations; those requiring internal fixation and re-manipulations are excluded)
• had a fracture manipulated in the emergency department
• received a follow-up outpatient appointment
• had a referral/appointment for physiotherapy
• had an X-ray in the emergency department

Full details of the coding used to identify these procedures/attendances is included in Appendix D.

In agreement with previous studies we find forearm fractures are the most common followed by elbow fractures, and also higher fractures incidences in the summer and older boys due to higher levels of physical activity^1,2. We find geographical variation in the recorded incidences of fractures in emergency care, some of this is likely the result of poor recording of fractures in emergency care in some areas. However, a previous paper found significant geographical variation with lower rates in London, similar to our results¹. It is thought that regional variation may be due to levels of physical activity, obesity levels and the ethnic make up of the area; fracture rates were found to be higher in white children compared to those of black or South Asian ethnicity and to correlate with obesity and levels of physical activity¹.

In the Paediatric Trauma and Orthopaedic Surgery GIRFT report it was suggested that there was scope to reduce the number of paediatric forearm manipulations in theatre by 80%³. That study used data from 2016-2019, in our study we have found since 2019/2020 the proportion of forearm fractures manipulated in theatre has fallen by 54%, from just over 8% in 2019/20 to under 4% in 2023/24.

It appears that there is the possibility to further reduce manipulations in theatre, particularly at some trusts, although overall at a national level the decrease in manipulation in theatre is likely to start slowing as the numbers get smaller. The switch to manipulating more fractures in the emergency department rather than in theatre will not only would save the NHS money, operating theatre time and resources, but will improve care for children and their families as treatment is quicker and less disruptive.

As highlighted within the Paediatric Trauma and Orthopaedic Surgery GIRFT report there are a number of factors that may be reduce the proportion of manipulations in the emergency department, including concerns about breach times, a lack of space/facilities to perform sedation, and a lack of familiarity with techniques³. Concerns around litigation may also lead to less conservative management, especially with elbow fractures which can lead to rare, but severe, complications. It is possible factors related to culture, lack of experience and litigation may also influence conservative management when it comes to reducing x-rays for clavicle and toe fractures and reducing follow-up appointments.

Nottingham University Hospitals NHS Trust, who are known to utilise more conservative management of paediatric fractures including routinely manipulating forearm fractures in the emergency department^3,4, have, as expected, one of the lowest proportions of forearm fractures manipulated in theatre. In 2022/2023 only 0.5% of forearm fractures were manipulated in theatre (while 11.81% were manipulated in the emergency department) meaning only 4% of all forearm fractures requiring manipulation were performed in theatre. In terms of follow-up appointments 57% of forearm fractures receive a follow-up appointment, which is just above the lowest quartile.

While at a national level the scope for significant further reductions in manipulations in theatre is likely to become limited, there is perhaps more scope in the future to reduce activity in terms of reducing unnecessary x-rays in the emergency department for clavicle and toe fractures, and in reducing the number of follow-up appointments.

The total number of forearm fractures that are manipulated has been decreasing, and there is perhaps scope for this to decrease further. Indeed, the CRAFFT study being conducted at over 45 sites in England is investigating whether there are any differences in outcomes between surgical reduction versus non-surgical casting for displaced distal radius fractures in children¹.

The GIRFT report found some trusts have moved towards splinting fractures in the emergency department with a virtual follow-up, rather than a temporary cast and then referral to a fracture clinic¹. From our data the use of virtual follow-ups appears widespread, but there may be scope to reduce virtual appointments, especially as some trusts with high overall follow-up rates have a particularly high number of virtual appointments.

Evidence suggests that many x-rays of clavicle fractures are unnecessary as they tend to not add significantly to a clinician’s assessment or alter management². However, almost all trusts still are x-raying the vast majority of potential clavicle fractures so the proportion of x-rays is still high even for the lowest decile of trusts (77%). Therefore our numbers are likely to be an underestimate of what activity savings could be potentially achieved. The same is likely true for toe fractures. It has been found that while most more experienced clinicians were comfortable treating clavicle fractures without x-rays, more junior clinicians were not².

Our absolute number estimates of reductions in activity may be an underestimate, because diagnoses are not reliably being recorded by some trusts in the emergency department. It is also possible that there may be some bias on the types of fractures attending emergency care that are given a diagnosis code. For example, it is possible at trusts that do not consistently record diagnosis codes in emergency care may be more likely to record fractures requiring manipulation than those that do not require manipulation or follow-up. However, when looking at trusts there was no significant relationship between the proportion of emergency care attendances without diagnosis codes and the proportion of fractures manipulated in theatre.

This report does highlight the need for more accurate and consistent recording of diagnoses in emergency care.

We are relying on the coding/reporting of fractures, manipulations, x-rays and outpatient appointments within the the SUS datasets. Some of the variability in the number of fractures, x-rays, physiotherapy referrals, follow-up appointments and manipulations between trusts and over time could be due to issues with the coding.

As diagnosis code is not always recorded in the emergency care dataset our incidence rates give the incidence of fractures recorded in emergency care, and only include closed fractures, open fractures were excluded.

The coding of fractures is not specific enough to determine at an individual level which fractures could be managed more conservatively, so we are relying on comparing proportion between trusts, but but we can’t be sure whether all trusts have a similar proportion of more complex fractures that do require manipulation and follow-up.

There will be a small number of cases where a child has more than one fracture or other injuries recorded in emergency care (<1%), so it is possible any follow-up appointments or manipulations could potentially be for a different injury/fracture sustained at the same time.

Data is allocated to the trust where the child attended the emergency department, e.g. if a child first attended the emergency department at trust A while on holiday but subsequently received follow-up appointments at trust B the fracture and subsequent follow-up would be included in the data for trust A.

The number of elbow manipulations recorded in theatre appears lower than would be expected, which may be due to coding issues. The lack of detailed coding of fractures makes it difficult to determine what is an elbow fracture and may mean that some are classified as forearm fractures, e.g. the SNOMED code for ‘Closed fracture of radius (disorder)’ is the most commonly used, and while we have classified these as forearm fractures, fractures of the proximal radius would be considered an elbow fracture. This may in part explain why we see a smaller number of elbow manipulations in theatre than would be expected, as they are included within forearm fractures. Also, HRG groupings for manipulations in theatre do not distinguish between forearm and elbow procedures, so it is not possible to determine from these which are forearm and which are elbow manipulations. We have only considered elbow manipulations in theatre that could be conducted in the emergency department, i.e. those without internal fixation such as wires or pins. Therefore our elbow manipulation number might also be affect by the propensity of individual trusts to use internal fixation for elbow fractures. We have also not included re-manipulations in our data.

These potential reductions in activity are for England only, but there is likely also the potential for activity savings in the other countries of the UK.

Cost savings are calculated from HRG prices and are only indicative estimates of potential cost savings.